Cents and shenshibility: The role of reward in talker-specific phonetic recalibration

Speech perception occurs in a dynamic environment and is potentially affected by many factors external to the speech signal, including listener motivation and the perceived value of talkers. It is plausible that listeners consider the value of individual talkers (e.g., a person that is important to the listener versus a passing stranger) when learning the idiosyncrasies of their speech, perhaps by directing their attention to the more “valuable” talker. In general, talker-specific phonetic recalibration has proven to be remarkably robust, leaving open how permeable it is to domain-general influences like attention and value-directed perceptual learning. The current study asked whether the magnitude of the talker-specific learning effect was modulated by an external reward intended to bias listeners towards one talker versus the other. Using an established paradigm, participants learned idiosyncratic pronunciations of the /s/ and /ʃ/ fricatives by two talkers—“Jane” and “Austin.” Over multiple experiments, participants were more likely to receive a few additional cents for one talker compared to the other following correct responses during exposure. We hypothesized that participants would show greater phonetic recalibration to the high-reward talker compared to the low-reward talker. Our results showed that participants did not show a difference in the degree of learning based on the value of each talker, suggesting that lexically guided perceptual learning is robust to external rewards. This finding opens discussions into the role of reward and attention in talker-specific phonetic recalibration, and how domain-general factors influence speech perception more broadly.